Clinical review

ABC of hypertension
Blood pressure measurement
Part IIConventional sphygmomanometry: technique of auscultatory blood
pressure measurement
Gareth Beevers, Gregory Y H Lip, Eoin OBrien

The measurement of blood pressure in clinical practice by the

century-old technique of Riva-Rocci/Korotkoff is dependent on This article has been adapted from the newly published 4th edition
the accurate transmission and interpretation of a signal of ABC of Hypertension. The book is available from the BMJ
(Korotkoff sound or pulse wave) from a subject via a device (the bookshop and at www.bmjbooks.com
sphygmomanometer) to an observer. Errors in measurement can
occur at each of these interactionary points of the technique, but
by far the most fallible component is the observer.

Observer error
In 1964, Geoffrey Rose and his colleagues classified observer
Rose classification of observer error
error into three categories.1
x Systematic error
Systematic error x Terminal digit preference
This leads to both intraobserver and interobserver error. It may x Observer prejudice
be caused by lack of concentration, poor hearing, confusion of
auditory and visual cues, etc. The most important factor is
failure to interpret the Korotkoff sounds accurately, especially
for diastolic pressure. Observer training techniques
Terminal digit preference x Direct instruction by an experienced observer
This refers to the phenomenon whereby the observer rounds off x Instruction manuals and booklets
the pressure reading to a digit of his or her choosing, most often x Audiotapes
to zero. Doctors may have a 12-fold bias in favour of the terminal x Video films
x CD Rom presentations
digit zero; this has grave implications for decisions on diagnosis
and treatment, although its greatest effect is in epidemiological
and research studies in which it can distort the frequency
distribution curve and reduce the power of statistical tests.2

Observer prejudice or bias

This is the practice whereby the observer simply adjusts the
pressure to meet his or her preconceived notion of what the
pressure should be. It usually occurs when there has been
recording of an excess of pressures below the cut-off point for
hypertension and it reflects the observers reluctance to
diagnose hypertension. This is most likely to occur when an
arbitrary division is applied between normal and high blood
pressure, for example 140/90 mm Hg. An observer might tend
to record a favourable measurement in a young healthy man
with a borderline increase in pressure, but categorise as
hypertensive an obese, middle aged man with a similar reading.
Likewise, there might be observer bias in overreading blood
pressure to facilitate recruitment for a research project, such as
a drug trial. Observer prejudice is a serious source of
inaccuracy, as the error cannot usually be demonstrated.3

Overcoming error by observer training

The technique of auscultatory blood pressure measurement is a
complicated one that is often taken for granted. Instruction to
medical students and nurses has not always been as
comprehensive as it might be, and assessment for competence
in measuring blood pressure has been a relatively recent
development.4 Ironically, these methods of achieving much
needed improvement in performing the auscultatory technique
have arrived as the mercury sphygmomanometer is under
threat and as automated devices move in to replace the Binaural stethoscope used for instruction in ausculatory blood pressure
observer; these have included: direct instruction using a measurement

BMJ VOLUME 322 28 APRIL 2001 bmj.com 1043

Clinical review

binaural stethoscope; the use of manuals, booklets, and

published recommendations; audiotape training methods;
videofilm methods, and, most recently, CD Rom methods. The
CD Rom produced by the Working Party on Blood Pressure
Measurement of the British Hypertension Society in 19985
incorporates instruction, with examples of blood pressure
measurement using a falling mercury column with Korotkoff
sounds and a means for the student to assess competence in the
technique using a series of examples. The CD is accompanied
by the British Hypertension Society booklet Blood pressure
measurement: recommendations of the British Hypertension Society.6
Overcoming error with instrumentation
As mentioned earlier, blood pressure measurement is subject to
observer prejudice and terminal digit preference, introducing
an error that is unacceptable for research work. Careful training
of observers can reduce but not abolish these sources of error,
some of which cannot be easily demonstrated. Because accuracy
of measurement is particularly desirable in research, efforts
have been made to devise devices that would minimise or
abolish observer error.
Measuring blood pressure
Assuming the observer has been trained and shown to be
proficient in the technique there are then a number of factors
that may affect the performance of the technique.5 6 Some of
these factors are described below.
Recommendations for observer training
Training observers in clinical practice: nursing and medical
students, doctors, paramedical personnel
x Instruction in the theory of hypertension and blood pressure
measurement
x Booklet for reading, eg BHS Recommendations on blood pressure
measurement
x Tutorial sessions with demonstrations using a binaural or
multiaural stethoscope
x CD Rom demonstration using, eg, the BHS CD Rom
x CD Rom assessment
x Repeat CD Rom assessment until level of accuracy achieved
x Reassessment using BHS CD Rom every two years
Training observers in research
x Measurement of blood pressurehighest possible standard
x Level of accuracy90% of SBP and DBP within 5 mm Hg
100% within 10 mm Hg of an expert observer
x Instruction in the theory of hypertension and blood pressure
measurement
x Audiogram to check auditory acuity
x Booklet for reading, eg BHS Recommendations on blood pressure
measurement
x Tutorial sessions with demonstrations using a binaural or
multiaural stethoscope
x CD Rom demonstration using, eg the BHS CD Rom
x CD Rom assessment
x Repeat CD Rom assessment until level of accuracy achieved
x Training and assessment repeated at least every three months

Attitude of observer
Before taking the blood pressure, the observer should be in a
comfortable and relaxed position, because if hurried the pressure
will be released too rapidly, resulting in underestimation of
systolic and overestimation of diastolic pressures. If any
interruption occurs the exact measurement may be forgotten and
an approximation made, so the blood pressure should always be
written down as soon as it has been measured.

Mercury and aneroid

sphygmomanometers
The mercury sphygmomanometer is a reliable device, but all too
often its continuing efficiency has been taken for granted,
whereas the aneroid manometer, which is not generally as Relaxed subject
accurate, is often assumed to be as reliable. These devices have
certain features in common; each has an inflation-deflation
system, and occluding bladder encased in a cuff, and both devices
measure blood pressure by auscultation using a stethoscope.6

Inflation-deflation system
The inflation-deflation system consists of an inflating and
deflating mechanism connected by rubber tubing to an
occluding bladder. The standard mercury and aneroid
sphygmomanometers used in clinical practice are operated
manually, with inflation being effected by means of a bulb
compressed by hand and deflation by means of a release valve,
which is also controlled by hand. The pump and control valve
are connected to the inflatable bladder and thence to the
sphygmomanometer by rubber tubing.

Rubber tubing
Leaks due to cracked or perished rubber make accurate
measurement of blood pressure difficult because the fall in
mercury cannot be controlled. The rubber should be in a good
condition and free from leaks. The minimum length of tubing
between the cuff and the manometer should be 70 cm and Mercury sphygmomanometer

1044 BMJ VOLUME 322 28 APRIL 2001 bmj.com


between the inflation source and the cuff the tubing should be
at least 30 cm in length. Connections should be airtight and
easily disconnected.
Control valve
A very common source of error in sphygmomanometers is the
control valve, especially when an air filter rather than a rubber
valve is used. Defective valves cause leakage, making control of
pressure release difficult; this leads to underestimation of
systolic and overestimation of diastolic pressures. Faults in the
control valve may be corrected easily by simply cleaning the
filter or replacing the control valve. It is helpful to have a
checklist of possible faults and the means of rectifying these.
Hazards of mercury
The mercury sphygmomanometer is a simple and accurate
device, which can be easily serviced, but there are rightly
concerns about the toxicity of mercury for individuals using
mercury sphygmomanometers, and for those who have to
service them. Users should be alert therefore to the hazards
associated with handling mercury.7
However, the greatest concern about mercury is its toxic
effects on the environment. The call to have mercury removed
from hospitals comes from the environmental lobby, which,
quite correctly, sees mercury as a toxic, persistent, and
bioaccumable substance. What happens, they ask, to the many
tons of mercury supplied for the manufacture of
sphygmomanometers and then distributed throughout the
world to hospitals and countless individual doctors? Quite
simply it finds its way back into the environment through
evaporation, sewage, or in solid waste, most seriously damaging
the marine environment, and it accumulates in soil and in
sediments thereby entering the food chain.
The mercury thermometer has been replaced in many
countries, and in Sweden and the Netherlands the use of mercury
is no longer permitted in hospitals. However, in other European
countries, including the UK and Ireland, the move to ban
mercury from hospital use has not been received with enthusiasm
on the grounds that there is no accurate alternative device to the
mercury sphygmomanometer. None the less, the fear of mercury
toxicity is making it difficult to get mercury sphygmomanometers
serviced, and the precautions recommended for dealing with a
mercury spill are influencing purchasing decisions. Indeed, this is
what central governmental policy in many countries would
favourthe gradual disappearance of mercury from hospitals
should a ban become operative.810
Preparing for the end of the mercury sphygmomanometer
Although it will be some years before any move is made to
replace the millimetre of mercury, we must prepare for changes
in clinical sphygmomanometry. Several simple measures can be
instigated immediately. Healthcare providers are being
encouraged to phase out mercury sphygmomanometers and
replace them only with devices that have been independently
validated against the relevant protocols. Automated devices
should provide blood pressures in both millimetres of mercury
and kilopascals, so that users can become familiar with
kilopascals. Finally, the medical and nursing professions, which
constitute the clinical market for blood pressure measuring
devices, must ensure that manufacturers provide us with accurate
devices designed to our specifications, rather than accepting, as
we have done in the past, devices in which these considerations
are secondary to the commercial success of the product.11
Aneroid manometers
Aneroid sphygmomanometers register pressure through a
bellows and lever system, which is mechanically more intricate
BMJ VOLUME 322 28 APRIL 2001 bmj.com
Clinical review
Consequences of defects in the control valve
Pumping control valve little or no effort required
Excessive squeeze on the pump filter blocked
With valve closed mercury at level steady
Falling mercury leak in inflation system
With valve released controlled fall of mercury
Failure to control mercury fall leak in inflation system
Advice to be included in the instructions accompanying a
sphygmomanometer using a mercury manometer
(from European Standard EN 1060-2)
B1 Guidelines and precautions
A mercury-type sphygmomanometer should be handled
with care. In particular, care should be taken to avoid
dropping the instrument or treating it in any way that could
result in damage to the manometer. Regular checks should
be made to ensure that there are no leaks from the inflation
system and to ensure that the manometer has not been
damaged so as to cause a loss of mercury.
B2 Health and safety when handling mercury
Exposure to mercury can have serious toxicological effects;
absorption of mercury results in neuropsychiatric disorders
and, in extreme cases, nephrosis. Therefore precautions
should be taken when carrying out any maintenance to a
mercury-type sphygmomanometer.
When cleaning or repairing the instrument, it should be
placed on a tray having a smooth, impervious surface which
slopes away from the operator at about 10 to the horizontal,
with a water filled trough at the rear. Suitable gloves (eg of
latex) should be worn to avoid direct skin contact. Work
should be carried out in a well ventilated area, and ingestion
and inhalation of the vapour should be avoided.
For more extensive repairs, the instrument should be
securely packed with adequate packing, sealed in a plastic
bag or container, and returned to a specialist repairer. It is
essential that a high standard of occupational hygiene is
maintained in premises where mercury containing
instruments are repaired. Chronic mercury absorption is
known to have occurred in individuals repairing
sphygmomanometers.
B3 Mercury spillage
When dealing with a mercury spillage, wear latex gloves.
Avoid prolonged inhalation of mercury vapour. Do not use
an open vacuum system to aid collection. Collect all the
small droplets of spilt mercury into one globule and
immediately transfer all the mercury into a container, which
should then be sealed.
After removal of as much of the mercury as practicable,
treat the contaminated surfaces with a wash composed of
equal parts of calcium hydroxide and powdered sulfur mixed
with water to form a thin paste. Apply this paste to all the
contaminated surfaces and allow to dry. After 24 h, remove the
paste and wash the surfaces with clean water. Allow to dry and
ventilate the area.
B4 Cleaning the manometer tube
To obtain the best results from a mercury-type
sphygmomanometer, the manometer tube should be cleaned
at regular intervals (eg under the recommended maintenance
schedule). This will ensure that the mercury can move up and
down the tube freely, and respond quickly to changes in
pressure in the cuff.
During cleaning, care should be taken to avoid the
contamination of clothing. Any material contaminated with
mercury should be sealed in a plastic bag before disposal in a
refuse receptacle.
1045
Clinical review

than the mercury reservoir and column. The jolts and bumps of
everyday use affect their accuracy; they lose accuracy over time,
usually leading to falsely low readings with the consequent Pointer
underestimation of blood pressure. They are therefore less Hairspring
accurate in use than mercury sphygmomanometers. When Scale
calibrated against a mercury sphygmomanometer a mean
difference of 3 mm Hg is considered to be acceptable; however, Linkage
58% of aneroid sphygmomanometers have been shown to have
errors greater than 4 mm Hg, with about one third of these
having errors higher than 7 mm Hg.12 Moreover, aneroid
sphygmomanometry is prone to all the problems of the
auscultatory technique, namely observer bias and terminal digit
preference.
Bellows

Position of manometer
The observer should take care when positioning the
manometer:
x The manometer should be no further than three feet (92 cm)
away so that the scale can be read easily. Pressure
x The mercury column should be vertical (some models are
Mechanism of an aneroid sphygmomanometer
designed with a tilt) and at eye levelthis is achieved most
effectively with stand mounted models, which can be easily
adjusted to suit the height of the observer.
x The mercury manometer has a vertical scale and errors will
occur unless the eye is kept close to the level of the meniscus.
The aneroid scale is a composite of vertical and horizontal
divisions and numbers, and must be viewed straight on with
the eye on a line perpendicular to the centre of the face of
the gauge.

Placing the cuff

The cuff should be wrapped around the arm ensuring that the
bladder dimensions are accurate. If the bladder does not
completely encircle the arm its centre must be over the brachial
artery. The rubber tubes from the bladder are usually placed
inferiorly, often at the site of the brachial artery, but it is now
recommended that they should be placed superiorly or, with
completely encircling bladders, posteriorly, so that the
Correct placement of cuff and bladder
antecubital fossa is easily accessible for auscultation. The lower
edge of the cuff should be 2-3 cm above the point of brachial
artery pulsation.

Palpatory estimation of blood pressure

The brachial artery should be palpated while the cuff is rapidly
inflated to about 30 mm Hg above the point where the pulse
disappears; the cuff is then slowly deflated, and the observer
notes the pressure at which the pulse reappears. This is the
approximate level of the systolic pressure. Palpatory estimation
is important because phase I sounds sometimes disappear as
pressure is reduced and reappear at a lower level (the
auscultatory gap), resulting in systolic pressure being
underestimated unless already determined by palpation. The
palpatory technique is useful in patients in whom auscultatory
endpoints may be difficult to judge accuratelyfor example,
pregnant women, patients in shock, or those taking exercise.
(The radial artery is often used for palpatory estimation of the
systolic pressure, but by using the brachial artery the observer
also establishes its location before auscultation.) Palpating artery

1046 BMJ VOLUME 322 28 APRIL 2001 bmj.com


Auscultatory measurement of systolic
and diastolic pressures
x Place the stethoscope gently over the brachial artery at the
point of maximal pulsation; a bell end-piece gives better
sound reproduction, but in clinical practice a diaphragm is
easier to secure with the fingers of one hand and covers a
larger area.
x The stethoscope should be held firmly and evenly but
without excessive pressuretoo much pressure might distort
the artery, producing sounds below diastolic pressure. The
stethoscope end-piece should not touch the clothing, cuff, or
rubber tubes to avoid friction sounds.
x The cuff should then be inflated rapidly to about 30 mm Hg
above the palpated systolic pressure and deflated at a rate of
2-3 mm Hg per pulse beat (or per second), during which the
auscultatory phenomena will be heard.
x When all sounds have disappeared the cuff should be
deflated rapidly and completely before repeating the
measurement to prevent venous congestion of the arm. The
phases shown in the box, which were first described by
Nicolai Korotkoff and later elaborated by Witold Ettinger, can
be heard.13
Diastolic dilemma
For many years recommendations on blood pressure
measurement have been uncertain about the diastolic
endpointthe so called diastolic dilemma. Phase IV (muffling)
may coincide with or be as much as 10 mm Hg higher than
phase V (disappearance), but usually the difference is less than
5 mm Hg; phase V correlates best with intra-arterial pressure.
There has been resistance to general acceptance of the silent
endpoint until recently, because the silent endpoint can be greatly
below the muffling of sounds in some groups of patients
children, pregnant women, anaemic or elderly patients. In some
patients sounds may even be audible when cuff pressure is
deflated to zero. There is now a general consensus that
disappearance of sounds (phase V) should be taken as diastolic
pressure except in those subjects mentioned above (as originally
recommended by Korotkoff in 1910).13
Recording blood pressure
The points to be noted when measuring blood pressure are
listed in the box opposite.
Number of measurements
One measurement should be taken carefully at each visit, with a
repeat measurement if there is uncertainty or distraction; do
not make a number of hurried measurements.
As a result of the variability of measurements of casual
blood pressure, decisions based on single measurements will
result in erroneous diagnosis and inappropriate management.
Reliability of measurements is improved if repeated
measurements are made. The alarm reaction to blood pressure
measurement may persist after several visits, so for patients in
whom sustained increases of blood pressures are being
assessed, a number of measurements should be made on
different occasions over a number of weeks or months before
diagnostic or management decisions are made.
BMJ 2001;322:1043-7
BMJ VOLUME 322 28 APRIL 2001 bmj.com
Clinical review
Auscultatory sounds
x Phase IThe first appearance of faint, repetitive, clear tapping
sounds which gradually increase in intensity for at least two
consecutive beats is the systolic blood pressure
x Phase IIA brief period may follow during which the sounds soften
and acquire a swishing quality
x Auscultatory gapIn some patients sounds may disappear altogether
for a short time
x Phase IIIThe return of sharper sounds, which become crisper to
regain, or even exceed, the intensity of phase I sounds. The clinical
significance, if any, to phases II and III has not been established
x Phase IVThe distinct abrupt muffling of sounds, which become
soft and blowing in quality
x Phase VThe point at which all sounds finally disappear completely
is the diastolic pressure
What to note when measuring blood pressure
x The blood pressure should be written down as soon as it has been
recorded
x Measurements of systolic and diastolic pressure should be made to
the nearest mm Hg
x Pressures should not be rounded off to the nearest 5 or
10 mm Hgdigit preference
x The arm in which the pressure is being recorded and the position
of the subject should be noted
x Pressures should be recorded in both arms on first attendance
x In obese patients the bladder size should be indicated
x If a standard cuff containing a bladder with the dimensions
23 12 cm has to be used, it is best to state this together with the
measurement so that the presence of cuff hypertension can be
taken into account in diagnostic and management decisions and
arrangements can be made for a more accurate measurement
x In clinical practice the diastolic pressure should be recorded as
phase V, except in those patients in whom sounds persist greatly
below muffling; this should be clearly indicated
x In hypertension research both phases IV and V should be recorded
x If the patient is anxious, restless, or distressed a note of this should
be made with the blood pressure
x The presence of an auscultatory gap should always be indicated
x In patients taking blood pressure lowering drugs the optimal time
for control of blood pressure will depend on the timing of the
drugs; when assessing the effect of antihypertensive drugs the time
of drug ingestion should be noted in relation to the time of
measurement
References
1 Rose G. Standardisation of observers in blood pressure measurement. Lancet
1965;1:673-4.
2 Keary L, Atkins N, Molloy E, Mee F, OBrien E. Terminal digit preference and
heaping in blood pressure measurement. J Hum Hypertens 1998;12:787-8.
3 OBrien E. Conventional blood pressure measurement. In: Birkenhager W, ed.
Practical management of hypertension. Dordrecht: Kluwer Academic Publishers, 1996:
13-22.
4 OBrien E, Mee F, Atkins N, OMalley K, Tam S. Training and assessment of observ-
ers for blood pressure measurement in hypertension research. J Hum Hypertens
1991;5:7-10.
5 The British Hypertension Society. Blood pressure measurement CD ROM. London:
BMJ Books, 1998.
6 OBrien E, Petrie J, Littler WA, de Swiet M, Padfield PD, Dillon MJ. Blood pressure
measurement: Recommendations of the British Hypertension Society. 3rd ed. London:
BMJ Books, 1997.
7 European Standard EN 1060-2 (British Standard BSSEN 1060-2: 1996). Specifica-
tion for non-invasive sphygmomanometers. Part 2. Supplementary requirements for
mechanical sphygmomanometers. 1995. European Commission for Standardisation,
Brussels.
8 OBrien E. Ave atque vale: the centenary of clinical sphygmomanometry. Lancet
1996;348:1569-70.
9 OBrien E. Will mercury manometers soon be obsolete? J Hum Hypertens
1995;9:933-4.
10 OBrien E. Replacing the mercury sphygmomanometer. BMJ 2000;320:815-16.
11 OBrien E, Owens P. Classic sphygomanometry: a fin de siecle reappraisal. In: Bulpitt
C, ed. Epidemiology of hypertension. Handbook of hypertension. Amsterdam: Elsevier,
2000:130-51.
12 Burke MJ, Towers HM, OMalley K, Fitzgerald D, OBrien E. Sphygmomanometers
in hospitals and family practice: problems and recommendations. BMJ 1982;
285:469-71.
13 OBrien E, Fitzgerald D. The history of indirect blood pressure measurement. In:
OBrien E, OMalley K, eds. Blood pressure measurement. Handbook of hypertension.
Amsterdam: Elsevier, 1991:1-54.
1047